Abstract
This volume brings together some of the most important material in linguistics and in the social studies which touch upon linguistics. The writings of Edward Sapir are invaluable for their complete grasp of linguistics, for their approach to language and culture and personality, for the wonderful working of data which they exhibit. We all know what a never-ending source of learning and delight this was to Sapir’s students and friends. Now it becomes available to those who would learn today, to whoever can appreciate both subtlety and independence of thought.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Notes
Page numbers refer directly to the volume under review, without specifying the particular article involved. I would like to call attention to Stanley S. Newman’s very interesting review of this book IJAL 17 (1951), 180–5, in which there is some explanation of Sapir’s unusual style of writing.
Ferdinand de Saussure, Cours de linguistique générale, 125.
Stanley S. Newman, Yokuts Language of California, New York 1944. (See Paper XII of this volume.)
Cf. Sapir’s article on glottalized continuants (225–50), and Henry M. Hoenigswald, ‘Sound Change and Linguistic Structure’, Lg. 22 (1946), 138–43.
To make this more explicit: Suppose all word-initial morphemes have two or more syllables (vowels). Then the probability of finding Ɂ rather than some other consonant after the FIRST vowel of a word is related simply to the frequency of the medial glottal stop. The probability of finding Ɂ after the Second vowel is related to the frequency of the glottal stop (medial and at the end of morphemes) plus the frequency of morphemes which end with a vowel (and of morphemes which begin with a vowel). However, the probability of finding other consonants (not Ɂ) after the second vowel is related merely to the frequency of those consonants medially and at morpheme-end.
How different Sapir’s psychologism is from this will be discussed in Part 3 below. For the moment, it is worth noting that Sapir’s grammatical formulations stayed within lin-guistic categories. In descriptive linguistics he would not say that people inserted a glottal stop so as to avoid the sequence VV, but that the glottal stop constituted, in respect of medial VV, a ‘protection’ (in cross-boundary position) of that non-occurrence of VV. The primacy of medial VV over the cross-boundary case is maintained, but in terms of the structure rather than in terms of people’s intervention in their own speech behavior.
We can say that the use of base forms in morphophonemics — as in Leonard Bloom-field’s ‘Menomini morphophonemics’, TCLP, 8 (1939), 105-15 — is a further step from history or process toward purely distributional statements.
It is interesting that Bloomfield’s work, which (as suggested above) represents a later stage in this particular development, presents phonemes no longer as the result of process but as direct classification, whereas the morphology is still largely described in terms of process. Cf. the chapters on phonology and on morphology in his book Language.
In Sapir and Swadesh, Nootka Texts Philadelphia 1939, 236-9.
As in Morris Swadesh and C. F. Voegelin, ‘A Problem in Phonological Alternation’, Lg. 15 (1939; written some years earlier), 7.
Sapir means: sounds as phonemically heard (perceived, structured) by the naive speaker and hearer.
Note ‘relational behavior’ for our ‘distribution’. The hearer might also classify it as a ‘bad pole’, so that even if the difference between the halfway sound and the regular sounds is noticed and not lost, it is nevertheless referred to (i. e. structured in terms of) the functionally (distributionally) determined points of the pattern.
We omit here the important difference that an English verb by itself contrasts most immediately with the small class of affix combinations (e. g. verb plus-ed), and only secondarily with a vast class of phrasal sequences in which that verb could be set (of which to cause to do so-and-so is one), while a Nootka verb by itself contrasts with a few specific combinations of verb plus affix (of which the causative affix is one), and only secondarily with the large class of phrasal sequences.
Hypnotism, 496 (where the Guthrie quotation is given in full) = Chap. 15 of J. McV. Hunt (ed.), Personality and the Behavior Disorders, Vol. I.
This was published in 1933. The novelty of this view may be seen from the fact that in 1929 Sapir had given it a more traditional formulation: “If I shove open a door in order to enter a house, the significance of the act lies precisely in its allowing me to make an easy entry. But if I ‘knock at the door’, a little reflection shows that the knock itself does not open the door for me. It serves merely as a sign that somebody is to come to open it for me” (163–4). His later understanding would suggest that the knock can be viewed instead as a tool, an indirect step in the course of getting the door opened (like the stick with which Köhler’s ape knocks down the banana, or the lever with which we pry up a rock). It is part of the continuous behavior which makes the person inside unlock the door for us, or which makes him ready for our intrusion. It is not a’ substitute for shoving’ but rather the equivalent for shoving in a society where people are customarily apprised of a visitor’s arrival. In social situations where this is not customary (as among intimates), one indeed opens the door without knocking.
For’ sign’ we should say: any associated behavior, such as a noise.
Note Martin Joos’s statement of it in the last paragraph of his paper ‘Description of Language Design’, Jour. Acoustic Soc. America 22 (1950), 707.
It is conceivable that there might have been yet another element of symbolism in language, if the noise behavior that became dissociated had had such a relation to the situation with which it was associated as would be independently arrived at by every speaker (or by every speaker in the given culture). Such associations occur in onomatopoetic elements (14), and they would have made words more a matter of individual expression than of arbitrary social learning. Sapir found some traces of such phonetic symbolism by a neat use of the methods of experimental psychology; part of this work appears in the present volume (61–72), part is as yet unpublished.
E. g. what large open classes there were (such as stems, or distinct verb and noun classes) which occurred with small closed classes (such as affixes, or distinct verb and noun affixes in various environmental subclasses).
It is always possible, of course, to overlook various environmental factors in analyzing the meanings of words. Sapir says (140): “If a quantitative goal is to be reached by increase, say ‘ten pages of reading’, more than necessarily has an approving ring (e. g., ‘I have already read more than three pages’, though it may actually be less than four), less than a disapproving ring (e. g., ‘I have only read less than eight pages’, though it may actually be more than seven). On the other hand, if the quantitative goal is to be reached by decrease, say ‘no more reading to do’, more than has a disapproving ring (e. g., ‘I have still more than three pages to do’, though actually less than four remain to be done), less than an approving ring (e. g., ‘I have less than eight pages to do’, though more than seven pages remain to be done out of a total of ten).” — If the form of the verb were taken into consideration here, it might be possible to show that the approving ring comes from the conjunction of more with the past tense and less with to do, the disapproving ring from more plus to do and less plus the past tense. To isolate the ‘affect in grading’, which Sapir seeks here, we extract an element ‘approval’ out of more plus past and less plus future, and an element ‘disapproval’ out of the opposite combinations.
As is well known, logic and especially semantics are also based in part upon the language of their practitioners, and are limited by their linguistic experience. However, this linguistic basis is not explicit because usually unacknowledged; narrow because usually limited to European languages; and arbitrary because not subject to explicit empirical and analytic techniques or to controls.
See E. Sapir and M. Swadesh, ‘American Indian Grammatical Categories’, Word 2, (1946), 103–12 — an item not included in the bibliography. On p. 111 Swadesh quotes a perfectly valid note of Sapir’s: “Naiveté of imagining that any analysis of experience is dependent on pattern expressed in language. Lack of case or other category no indication of lack functionally.... In any given context involving use of language, lang, response is not to be split up into its elements grammatically nor sensorimotorly but kept as unit in contextual pattern.” Elsewhere, however, Sapir says: “The ‘real world’ is to a large extent unconsciously built up on the language habits of the group... The worlds in which different societies live are distinct worlds, not merely the same world with different labels attached” (162). There is no contradiction here, since the ‘environing world’ is the physical world, whereas the ‘real world’, in quotes, is also called’ social reality’ (162) and constitutes the physical world as socially perceived: “Even the simplest environmental influence is either supported or transformed by social forces” (89); “The physical environment is reflected in language only insofar as it has been influenced by social forces” (90).
Norbert Wiener, Cybernetics, Chap. 3, esp. 75-9.
For an example of how particular logical relations can be built into a constructed language, consider the ‘newspeak’ of George Orwell’s novel Nineteen eighty-four. One of the distributional features which is only lightly suggested in his system is the technique (not unknown in our real languages) of letting opposites equal or replace each other in certain environments, with the result that no distinction between opposites (say between war and peace) can be made in the language.
Cf. Zellig S. Harris, Development of the Canaanite Dialects, New Haven 1939, 99-100.
An echo of this appears in the work of Sapir’s student Morris Swadesh on rate of vocabulary change. Cf. in particular his Salish investigations, carried out under the auspices of the Boas Collection in the American Philosophical Society Library, and published in’ salish Internal Relationships’, IJAL 16 (1950), 157-67.
His writings on music dealt with its relation to poetry rather than to linguistic structure. See for example his article ‘The Musical Foundations of Verse’, JEGP 20 (1921), 213-28.
Comparison of the page references in this paragraph with those in the preceding one will indicate which point is included in which category. Thus, simplicity of form is discussed under the seriation of culture elements from the simple or primary to the derived.
But one must guard against such other factors as simplification of a form in the course of borrowing (402).
But a cultural element (e. g. ritual use of tobacco) may be borrowed without its chronological antecedent (cultivation of the tobacco plant; 403).
Larger areas will often not mean greater age, since some elements diffuse faster than others (e. g. elements which are not secret, or are detachable from their context, 414–5), and some environments favor quicker diffusion (e. g. areas covered by related languages, or lines of easy communication, 416–9).
But some languages favor descriptive word formations (437); and old non-descriptive words may have been changed in meaning to apply to a later culture element (438).
And languages differ in their hospitality to analogic regularization of grammar (442).
In language, unlike culture, borrowed material can often be readily recognized by its phonetic structure, morphological unanalyzability, length, or the like, and can be traced to its language of origin (445–9).
In discussing what can be learned historically from the way a language family is spread and diversified (453–8), Sapir says that the fact that both Aleut and Eskimo are spoken in Alaska, while only Eskimo is spoken in Canada, supports Alaska as the center of dispersion of Eskimo. But such considerations will not hold if there are successive waves of emigration from a center, which pile up at coast-lines or other boundaries, thus making the periphery more differentiated in language than the center (cf. the diversified Semitic periphery as against the Arabian center).
In some cases, as in social organization or linguistic usage and vocabulary, the individual carries out only a part of the socially observed pattern (516), and we cannot say that his selection of behavior is the same as the social pattern. In other cases, as in grammatical structure, the individual’s behavior is virtually the same as that which is described for the society as a whole.
E. g. of ‘explaining’ an unusual suffix by analyzing it as a combination of two suffixes which are members of classes whose sequence would indeed occur precisely in the position occupied by the strange suffix.
In linguistics, the analog to “the way personalities enter into relations with each other” is the distributional interrelation of elements.
Note also the ‘feeling’ due to the range of occurrences of the morpheme-ké·h in Navaho (220–2).
A descriptive pattern would have been the same as a configurational pressure on the individual speakers; a historical reconstruction would not.
Sapir seems to say that the native ‘grasps’ the social pattern, while the outside observer just sees the resultant behavior (547). But by observing enough of the behavior, the observer can see as much as the native has grasped. The native himself has grasped it only by observing a great deal of behavior; he is a ‘participant observer’ of his own society. Hence the social patterns are really not ‘felt’ by him, but observed; the observations are experiences upon which he can draw when he acts.
Gordon Childe, What Happened in History, 8: “In practice ideas form as effective an element in the environment of any human society as do mountains, trees, animals, the weather and the rest of external nature. Societies, that is, behave as if they were reacting to a spiritual environment as well as to a material environment.”
The siren of literary effect, which is not always identical with meticulous statement, sometimes led Sapir into such sentences as this (106): “It is largely the function of the artist to make articulate these more subtle intentions of society.” Some writers really mean it when they refer to the’ subtle intentions of society’ Sapir obviously did not.
The ‘higher solution’ lies in the fact that the Indian makes fuller use of available technical knowledge, and has more opportunity to participate in the arranging of his own work, and to make any changes in it. If he fails to carry out any developments in his own work, it is because he lacks the immediate need or the means to make such changes, not so much because he is restricted as to his own activities (as he would be in our society) by orders from others and by a tight organizational structure into which he fits as a cog. One might argue that not only our society has such restrictions, and that the problem is not so much whether primitive societies are freer of them but rather how in our society the people who work can become more free of such limitations. But Sapir’s comment is more important as a critique of his own culture than as a commendation of the Indian’s — which is natural, since Sapir knew the detailed difficulties of his own culture better, and since these touched him more closely.
In two somewhat earlier articles, Sapir talked in a more standard anthropological manner without recognizing the limitations of cultural uniformity that were stressed in the excerpt quoted above: “[A Haida Indian] cannot be born, become of age, be married, give feasts, be invited to a feast, take or give a name, decorate his belongings, or die as a mere individual, but always as one who shares in the traditions and usages that go with the Killer-whale or associated crests” (345). “If we leave the more sophisticated peoples and study the social habits of primitive and barbaric folk, we shall find that it is very difficult to discover religious institutions that are as highly formalized as those that go under the name of the Roman Catholic Church or of Judaism. Yet religion in some sense is everywhere present. It seems to be as universal as speech itself and the use of material tools” (346). Sapir’s own argument above leaves little room for doubt that we could find many actions by Haida individuals which would manage to keep clear of any crest identification, just as the second John Doe avoided involvement in the ward system; and that many primitive individuals are free of any religious identification, just as are many moderns. It is quite understandable that Sapir should have noticed the individual differences in his own society and missed them — or simply not had the data — in other societies which he necessarily knew in far less detail. When anthropologists have turned to write about their own society, they have customarily found conformity and acceptable conditions, in contrast with the class controls and major cultural inadequacies which they found in enemy or primitive societies. It is to Sapir’s credit that he used his critical powers where they might do most good: in remarks about his own society. He might omit some individual variation or some cultural critique of a primitive society, but he would be sure to fight it out at home. (See § 3.3, end, and Part 4 below.)
Sapir sometimes spoke of language as determining people’s perception: “The ‘real world’ is to a large extent unconsciously built up on the language habits of the group” (162). In later work he says rather that a particular language can “help and retard us in our exploration of experience” (11) — i. e. retard us but not stamp us irrevocably; that mathematics has gone on to develop its own alternative system (118), rather than remaining blocked by perceptions based on language; and that “as our scientific experience grows we must learn to fight the implications of language” (10), rather than accept it as inevitable that we can do no more than reflect the existing linguistic structure.
What the anthropologist constructs are cultural patterns. What members of the society observe, or impose upon others, are culturally patterned behaviors.
Princeton University official register 1941 § 33.3 (257): “Freshmen and sophomores are required to attend at least one-half of the Sunday services in the University Chapel in each quarter of the academic year.... To be appreciated, the service of public worship must be experienced and this is the basis of the requirement for chapel.” One can imagine what communication Sapir would have recognized when the regents of the university whose press published this volume of his selected writings demanded oaths of its faculty and fired the non-submissive.
Sapir adds: “We have said nothing so far that is not utterly commonplace. What is strange is that the ultimate importance of these commonplaces seems not to be thoroughly grasped by social scientists at the present time”. The strangeness disappears, of course, when one remembers that the social scientists are not catering to the rebels. As John F. Embree says (American Anthropologist NS 52 [1950], 431), “The applied anthropologist... advises managers how to manage their workers; he has been little concerned to advise the managed how to maintain their own social interests vis-a-vis the managers”.
Sapir, ‘Do we need a “Superorganic”?’, American Anthropologist NS 19 (1917), 444. This article is not included in the present volume.
This effect of the formal configuration may be seen most readily in limited well-structured fields, such as music or some particular science. Aside from the more generally social factors that lead to particular developments and tendencies in each field, it seems probable that the existing pattern at any one time (the kind of scales used, the type of composition potentialities which have been well investigated) favor certain directions of change, rather than others, by those who try out changes.
The effect of culturally patterned interpersonal relations is treated by Harry Stack Sullivan, Conceptions of Modern Psychiatry.
Cf. John F. Embree, ‘A Note on Ethnocentrism in Anthropology’, American Anthropologist NS 52 (1950), 430–2.
References
Selected Writings of Edward Sapir in Language, Culture, and Personality (Edited by David G. Mandelbaum), University of California Press, Berkeley and Los Angeles, 1949.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1970 Springer Science+Business Media Dordrecht
About this chapter
Cite this chapter
Harris, Z.S. (1970). Sapir’s Selected Writings . In: Papers in Structural and Transformational Linguistics. Formal Linguistics Series. Springer, Dordrecht. https://doi.org/10.1007/978-94-017-6059-1_33
Download citation
DOI: https://doi.org/10.1007/978-94-017-6059-1_33
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-017-5716-4
Online ISBN: 978-94-017-6059-1
eBook Packages: Springer Book Archive